A COMPREHENSIVE STUDY OF THE EAGLE FORD SHALE FRACTURE CONDUCTIVITY
A Thesis by
OMAR ENRIQUEZ TENORIO
Submitted to the Office of Graduate and Professional Studies of Texas A&M University
in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE
Chair of Committee, Ding Zhu Committee Members, A. Daniel Hill
Michael Pope Head of Department, A. Daniel Hill
August 2016
Major Subject: Petroleum Engineering
ii ABSTRACT
The aim of this study was to develop an understanding of the Eagle Ford shale unpropped and propped fracture conductivity. Samples were collected at Antonio Creek and Lozier Canyon. Both outcrops are several hundred feet thick and extend for a few thousand feet laterally. At these locations five different geological facies (A, B, C, D and E), each with distinct lithological characteristics and geochemical properties that can be correlated to the Eagle Ford shale in the subsurface are accessible. The mineralogy of the collected samples was determined via x-ray diffraction which corroborates the relationship of the samples to the outcrops and hence to the Eagle Ford shale in the subsurface.
After collected, the samples were cut into modify API conductivity cell dimensions. The fracture conductivity at different closure stresses was determined based on laboratory measurements of flow rate and pressure drops along the fracture. The fluid used in this work was nitrogen. The proppant concentration used is representative of what is commonly used in hydraulic fracturing treatments in the Eagle Ford shale. The heterogeneity of the Eagle Ford shale is addressed by obtaining conductivity samples in two different directions with respect to the bedding plane for the five geological facies. Three directions were obtained for facies B.
Fracture area and fracture roughness (surface attributes) were calculated from the data taken by a surface profilometer. The mechanical properties, namely Young’s modulus, Poisson’s ratio and Brinell hardness, were experimentally determined from core plugs acquired from the same rocks where the conductivity samples were obtained. The
iii
effect that the surfaces attributes, mechanical properties and geological facies (mineralogy) have on conductivity was analyzed and the major findings of this work are the following.
From this study, we found that fracture orientation has an impact on the fracture conductivity. It was also found that the geological characteristics (mineralogy) and mechanical properties (Poisson’s Ratio) impact the fracture conductivity. The geological facies (lithology) impact the fracture conductivity in Eagle Ford shale. Facies A, B and C exhibit a good relationship between fracture conductivity and surface attributes (fracture roughness and area). This relationship is not present in facies D and E.
This work provides a foundation for future studies (damage mechanism) of the Eagle Ford shale fracture conductivity and gives an insight into the relationship of fracture conductivity and geological facies.
iv
DEDICATION
I would like to dedicate this work to my family: My wife Sarah Henry, my parents Roberto Enriquez and Paola Tenorio, my siblings Edgar Enriquez and Paola Enriquez and to the rest of my family. Even though you may not know exactly what I am up to, I have always felt your unconditional support.
v
ACKNOWLEDGEMENTS
I would like to express my sincere gratitude to both Dr. Ding Zhu and Dr. Daniel Hill for giving the opportunity to join the fracture conductivity research group. The discussions and conversations that we have during weekly meeting and during individual meetings definitely enhance my professional development as well as my personal development. I would like to especially thank Dr. Zhu, for helping me to make the decision that graduate school was the right path for myself, after completing my studies I am confident that I chose the right path.
I would also like to thank Dr. Michael Pope for serving as a committee member. Next, I would like to thank BP for allowing the collection of samples in Lozier and Antonio creek canyons.
I would like to thank the members of the fracture conductivity research group that I have had the privilege to work with: Bill Foran, Dante Guerra, Cody Kainer, Ashley Knorr, Mark McGinley and Paola Perez. I can certainly say that I have learned something from all of you. I would like to thank Jesse Guerra for developing the code for the surface attributes. I would also like to thank all the previous members of the conductivity group, your work provided a solid foundation to start my own work.
I would like to thank John Maldonado and the rest of the Petroleum Engineering staff for their help during my experiments. Finally, I would like to thank RPSEA for the financial support provided for this work.
vi
NOMENCLATURE
𝐴𝑓𝑙𝑜𝑤 Cross sectional area of flow, L2, in2 𝐴𝑓𝑟𝑎𝑐𝑡𝑢𝑟𝑒 Fracture Area, L2, in2
bbl Barrel of oil, L3 , 42 US gal BI Brittleness index, dimensionless
𝐶𝑎 Areal proppant concentration, ML−2, lbm/ft2
cf cubic foot, L3
𝐶𝑓𝑑 Dimensionless fracture conductivity, dimensionless 𝐶𝑓 Fracture Conductivity, L2L, md-ft 𝑑𝑝 Proppant diameter, L, mm ℎ𝑓 Sample width, L, ft 𝑘𝑓 Fracture Permeability, L2, md-ft 𝑘𝑚 Matrix Permeability, L2, md-ft L Sample length, L, ft
𝑚̇ Mass flow rate, Mt−1, kg/min
𝑀𝑔 Molecular Mass, MM−1N−1, Kg/Kg mol
M One thousand, dimensionless
MM One Million, dimensionless
𝑃𝑐𝑒𝑙𝑙 Cell pressure, ML−1r2, psig
vii
Ppg Pounds per gallon, ML−3, lbm/gal P1 Upstream pressure, ML−1r2, psig q Volumetric flow rate, L3t−1, L/min
R Universal gas constant, ML2t−2N−1φ−1, J/mol-K
T Temperature, K
𝑤𝑓 Fracture width, L, in
v Fluid velocity, Lt−1, ft/s 𝑥𝑓 Fracture half length, L, ft
Z Gas compressibility factor, dimensionless
Greek
𝛾𝑝 Proppant specific gravity, dimensionless
∅ Porosity, fraction
𝜇 Fluid viscosity, ML−1T−2T, Pa-s ∆𝑃 Differential pressure, ML−1r2, psig 𝜌𝑓 Fluid density, ML−3, lbm/ft3
viii TABLE OF CONTENTS Page ABSTRACT ...ii DEDICATION ...iv ACKNOWLEDGEMENTS ...v NOMENCLATURE ...vi
TABLE OF CONTENTS ...viii
LIST OF FIGURES ...x
LIST OF TABLES ...xv
1 INTRODUCTION ...1
1.1 Background ...1
1.2 Literature Review ...2
1.2.1 Hydraulic Fracturing of Unconventional Reservoirs- Eagle Ford Shale ...2
1.2.2 Fracture Conductivity ...4
1.2.3 Fracture Conductivity Controlling Parameters ...6
1.2.4 Eagle Ford Depositional Environment and Lithofacies ...11
1.3 Problem Description, Objectives and Significance ...13
1.4 Approach ...15
1.5 Structure of Thesis ...16
2 EXPERIMENTAL DESING AND METHODOLOGY ...18
2.1 Sample Collection ...18
2.2 Sample Preparation for Fracture Conductivity Test ...26
2.3 Proppant Concentration ...28
2.4 Laboratory Procedures ...34
2.4.1 Apparatus ...35
2.4.2 Experimental Design ...40
2.4.3 Methodology of Surface Roughness and Surface Area Measurements ...42
ix
2.4.4 Methodology of Sample Preparation ...45
2.4.5 Methodology of Gas Fracture Conductivity Determination ...49
3 FRACTURE CONDUCTIVITY ON FACIES A-E ...57
3.1 Facies A ...57
3.1.1 Unpropped Fracture Conductivity Facies A ...58
3.1.2 Propped Fracture Conductivity Facies A ...60
3.2 Facies B ...62
3.2.1 Unpropped Fracture Conductivity Facies B ...64
3.2.2 Propped Fracture Conductivity Facies B ...65
3.3 Facies C ...67
3.3.1 Unpropped Fracture Conductivity Facies C ...68
3.3.2 Propped Fracture Conductivity Facies C ...68
3.4 Facies D ...70
3.4.1 Unpropped Fracture Conductivity Facies D ...71
3.4.2 Propped Fracture Conductivity Facies D ...71
3.5 Facies E ...73
3.5.1 Unpropped Fracture Conductivity Facies E ...74
3.5.2 Propped Fracture Conductivity Facies E ...75
4 FRACTURE CONDUCTIVITY COMPARISONS ACROSS FACIES A-E ...78
4.1 Fracture Conductivity Orientation ...78
4.2 Mechanical Properties and Brittleness Index ...80
5 CONCLUSIONS AND RECOMMENDATIONS ...87
5.1 Conclusions ...87
5.2 Recommendations and Limitations ...89
x
LIST OF FIGURES
Page Fig. 1.1 ‒ Drilling Permits and Average Daily Oil and Gas Production in the
Eagle Ford Shale (RRC 2016b)..………...3 Fig. 1.2 ‒Equivalent Skin Factor, Fracture Half Length and Dimensionless Fracture
Conductivity (Cinco-Ley et al 1978)...4
Fig. 1.3 ‒ Proppant Arrangement with Proppant Pillars and Channels on a Fracture Conductivity Sample (Gillard et al. 2010)……….7
Fig. 1.4 ‒ Major Structural Features of Eagle Ford Shale (Tian et al. 2012)…………...12 Fig. 1.5 ‒ Total Organic Content of the Lower Eagle Ford Shale Calculated from
Sonic and Resistivity Logs (Tian et al. 2013)………..………13 Fig. 1.6 ‒ Gas Oil Ratio for Black Oil Wells in the Eagle Ford Shale
(Tian et al. 2013)...14 Fig. 2.1 ‒ Map of the Geographic Location of Antonio Creek and Lozier Canyon
(Gardner et al. 2013)………19 Fig. 2.2 ‒ Five Different Geological Facies of the Eagle Ford Shale with Their
Corresponding Gamma Ray Response and the Industry Nomenclature for the Formation- at Lozier Canyon, Southwest Texas (Donovan et al.
2012)………..………..19 Fig. 2.3 ‒ Mineralogy (Weight Percent) and Collected Rock Sample
(Sharpie for Scale) – Facies A……….………20 Fig. 2.4 ‒ Mineralogy (Weight Percent) and Collected Rock Sample
(Pen for Scale) – Facies B………...………….21 Fig. 2.5 ‒ Mineralogy (Weight Percent) and Collected Rock Sample
(Credit Card for Scale) – Facies C………...22 Fig. 2.6 ‒ Mineralogy (Weight Percent) and Collected Rock Sample
(Pen for Scale) – Facies D………...22 Fig. 2.7 ‒ Mineralogy (Weight Percent) and Collected Rock Sample
xi
Fig. 2.8 ‒ Biostratigraphic, Geochemical, Petrophysical and Lithofacies Interpretation of Lozier Canyon and Sequence Interpretation
(Donovan et al. 2012)………..24
Fig. 2.9 ‒ Well Log Correlation from the Maverick Basin to the San Marcos Arch Modified from Donovan et al. (2012)………...…25
Fig. 2.10 ‒ Orientation of Conductivity Samples Obtained from the Outcrops ...…...…26
Fig. 2.11 ‒Modified API Conductivity Sample Dimensions (Kamenov 2013)………...27
Fig. 2.12 ‒ Sample Orientation: Parallel (X0) and Perpendicular (X90) to the Bedding Plane and on the Bedding plane (Z) Modified from McGinley 2015...28
Fig. 2.13 ‒ Proppant Concentration (Ppg) Histogram for 353 Wells Located in the Eagle Ford Shale in 16 Different Counties in Texas- Data Obtained from www. Drillinginfo.com………...……..31
Fig. 2.14 ‒ Proppant Type Histogram for 284 Wells Located in the Eagle Ford Shale in 16 Different Counties in Texas- Data Obtained from www. Drillinginfo.com……….32
Fig. 2.15 ‒ Proppant Concentration (Ppg) Ignoring and Taking into Account the Proppant Volume for Multiple Proppant (Grain) Densities…...…………...34
Fig. 2.16 ‒ Process Flow Chart to Calculate and Analyzed the Eagle Ford Shale Fracture Conductivity………..35
Fig. 2.17 ‒ Modified API Conductivity Cell Modified from Kamenov 2013.…….……37
Fig. 2.18 ‒ Low and High Flow Rates Mass Flow Meters………...38
Fig. 2.19 ‒ Pressure Transducers Validyne DP 15 (Validyne Engineering 2016).…...38
Fig. 2.20 ‒ Schematic of Fracture Conductivity Experimental Setup………..39
Fig. 2.21 ‒ Surface Profilometer..……….…39
xii
Fig. 2.23 ‒ Fracture Area and Roughness from Profilometer Measurements……. ....…44 Fig. 2.24 ‒ Aligned Conductivity Sample in the Sample Preparation
Mold (Guzek 2014)………...45 Fig. 2.25 ‒ Removing the Conductivity Sample Coated With Epoxy from
the Sample Preparation Mold………46 Fig. 2.26 ‒ Conductivity Sample Ready to be Inserted into the Modify API
Conductivity Cell for Measurements……….47 Fig. 2.27 ‒ Cutting the Epoxy from the Conductivity Sample for
Proppant Placement...48 Fig. 2.28 ‒ Conductivity Sample with 0.1 lbm/ft2 100 Mesh...49 Fig. 2.29 ‒ Fully Assemble Fracture Modified API Conductivity Cell Front
and Back Views...50 Fig. 2.30 ‒ Fracture Conductivity via Darcy’s Law for Sample EE_X90_2_100
at 1000 Psi Closure Stress...55 Fig. 2.31 ‒ Fracture Conductivity via Darcy’s Law for Sample EE_X90_2_100...56 Fig. 3.1 ‒ Unpropped Fracture Conductivity, Fracture Area and
Facies A………..…58 Fig. 3.2 ‒ Samples EA_X0_1 and EA_X0_2 Weak Planes in the Fracture Faces…...…59 Fig. 3.3 ‒ Exponential Decline of Unpropped Fracture Conductivity-Facies A………..60 Fig. 3.4 ‒ 100 Mesh- 0.1 lbm/ft2 Fracture Conductivity, Fracture Area and
Roughness-Facies A……….………….………...…..61 Fig. 3.5 ‒ 30/50 Mesh- 0.1 lbm/ft2, 100 mesh- 0.1 lbm/ft2and Unpropped
Fracture Conductivity- Sample: EA_X90_2……...…...….……….….62 Fig. 3.6 ‒ Facies B: Dark to Gray Samples-Sample: EB_Z_1……….……63 Fig. 3.7 ‒ Average Unpropped Fracture Conductivity, Exponential Decline Trend
xiii
Fig. 3.8 ‒ 100 Mesh-0.l bm/ft2Fracture Conductivity, Exponential Decline
Trend Line and Fracture Area and Roughness-Facies B……….…...65 Fig. 3.9 ‒ 30/50 Mesh-0.1 lbm/ft2Fracture Conductivity, Exponential Decline
Trend Line and Fracture Area and Roughness-Facies B….………...…66 Fig. 3.10 ‒ 30/50 Mesh-0.1 lbm/ft2, 100 Mesh-0.1 lbm/ft2and Unpropped
Fracture Conductivity X90 Direction-Facies B……...………...….66 Fig. 3.11 ‒ Facies C-Sample EC_X90_1………….……….67 Fig. 3.12 ‒ Unpropped Fracture Conductivity Sample: EC_X90_1-Facies C…..……...68 Fig. 3.13 ‒ 100 Mesh-0.1 lbm/ft2Fracture Conductivity, Exponential Decline
Trend Line and Fracture Area and Roughness-Facies C……….……...…69 Fig. 3.14 ‒ 30/50 Mesh-0.1 lbm/ft2, 100 Mesh-0.1 lbm/ft2 and Unpropped
Fracture Conductivity-Facies C……….………...…69 Fig. 3.15 ‒ Facies C-Sample EC_X90_1……….……….70 Fig. 3.16 ‒ Unpropped Fracture Conductivity, Exponential Decline Trend Line
and Fracture Area and Roughness-Facies D…….……….……71 Fig. 3.17 ‒ 100 Mesh-0.1 lbm/ft2Fracture Conductivity, Exponential Decline
Trend Line and Fracture Area and Roughness-Facies D…….………..72 Fig. 3.18 ‒ 30/50 Mesh-0.1 lbm/ft2, 100 Mesh-0.1 lbm/ft2 and Unpropped
Fracture Conductivity-Facies D…….…...………..….73 Fig. 3.19 ‒ Facies E-Sample EE_X90_1……….……….74 Fig. 3.20 ‒ Unpropped Fracture Conductivity, Exponential Decline Trend Line
and Fracture Area and Roughness-Facies E….………....75 Fig. 3.21 ‒ 100 Mesh-0.1 lbm/ft2 Fracture Conductivity, Exponential Decline
Trend Line and Fracture Area and Roughness-Facies C……….…………..76 Fig. 3.22 ‒ 30/50 Mesh-0.1 lbm/ft2 Fracture Conductivity, Exponential Decline Trend
Line and Fracture Area and Roughness-Facies E…….……...…...76 Fig. 3.23 ‒ 30/50 Mesh-0.1 lbm/ft2, 100 Mesh-0.1 lbm/ft2 and Unpropped
xiv
Fig. 4.1 ‒ Fracture Conductivity for 100 Mesh-0.1 lbm/ft2 Concentration
in the X0 and X90 Directions for Facies A, B and C……….……78 Fig. 4.2 ‒ Fracture Conductivity for 100 Mesh-0.1 lbm/ft2 Concentration
in the X0 and X90 Directions for Facies D and E………...…79 Fig. 4.3 ‒ Fracture Conductivity for 100 Mesh-0.1 lbm/ft2 Concentration
Facies B: Samples in Z, X0 and X90 Directions and Fracture
Area and Roughness………...….80 Fig. 4.4 ‒ Fracture Conductivity Facies A-E and Brittleness Indexes. 100 Mesh-
0.1 lbm/ft2 Concentration…………...……….83
Fig. 4.5 ‒ Fracture Conductivity Facies A-E and Brittleness Indexes. 30/50 Mesh-
0.1 lbm/ft2 Concentration…………..………...……….…85 Fig. 4.6 ‒ Fracture Conductivity Facies A-E and Brittleness Indexes-
xv
LIST OF TABLES
Page Table. 2.1 ‒ Total Organic Content (Weight Percent) Samples from Lozier Canyon.
Sub-Facies B1 and B2 Combined into Facies B…………...…………...23 Table. 2.2 ‒ Equivalent Proppant Concentrations (Ppg) for Multiple Combinations
of Fracture Widths (in) and Proppant Concentrations per
Unit Area lbm/ft2...…………..….……….……….30 Table. 2.3 ‒ Conductivity Samples Obtained from the Collected Eagle Ford
Shale Outcrop Rocks………....40 Table. 2.4 ‒ Performed Fracture Conductivity Experiments for the Eagle Ford Shale
Samples………...…………..……….42 Table. 3.1 ‒ Surface Attributes for the Different Experimental Conditions
Facies A………...57 Table. 3.2 ‒ Surface Attributes for the Different Experimental Conditions
Facies B………...…63 Table. 3.3 ‒ Surface Attributes for the Different Experimental Conditions
Facies C………..….67 Table. 3.4 ‒ Surface Attributes for the Different Experimental Conditions
Facies D………..….70 Table. 3.5 ‒ Surface Attributes for the Different Experimental Conditions
Facies E………....74 Table. 4.1 ‒ Summary of the Mechanical Properties: Young’s Modulus, Poisson
Ratio and Brinell Hardness for the Eagle Ford Shale Samples…………...81 Table. 4.2 ‒ Brittleness Index for the Eagle Ford Shale Samples………82 Table. 4.3 ‒ Number of Data Points Used to Calculate the X0 and X90
1
1. INTRODUCTION
1.1 Background
Hydraulic fracturing is a stimulation process that involves pumping a pressurized mixture of fluids, chemicals and proppants (usually sand) down a well. On high permeability reservoirs the objective of hydraulic fracturing is to bypass a damage zone near the wellbore. On the other hand, in low permeability formations like shale reservoirs the objective is to increase the contact area with the reservoir so commercial flow rates can be obtained. The hydraulic fracturing process begins with pumping pad fluid at pressures higher than the formation fracturing pressure which fractures the rock. Next, proppants are added to the fluids and it is expected that the proppants will fill the newly created fractures and keep them open once the pumping stage is over. Chemicals are added at every stage of the hydraulic fracturing process. These chemicals are used to reduce the friction between the slurry and the pipe, increase and maintain the fluid viscosity, limit corrosion and scale deposition in the pipe, clean the perforations and to avoid/minimize formation damage.
Unconventional reservoirs, specifically shale reservoirs, are heterogeneous as can be seen from the wide range of productions that are observed from the same formation. For example, the Eagle Ford shale produces dry gas on its eastern regions and black oil on its western region. Production rates and fluid characteristics from wells located in the same region can be very different. Therefore understanding the impact that reservoir heterogeneity has on the fracture conductivity is essential to understand shale reservoirs.
2
With this in mind we performed a systematic investigation of the fracture conductivity of multiple lithological facies presented in the Eagle Ford shale.
1.2 Literature Review
In this section a literature review of the fracture conductivity, its controlling parameters, the Eagle Ford shale and its importance are presented.
1.2.1 Hydraulic Fracturing of Unconventional Reservoirs-Eagle Ford Shale
The first experiments of the oil and gas industry with hydraulic fracturing were performed more than sixty years ago. By 1950 the process was commercialized and horizontal wells were common by the late 1970s (King 2012). Modern hydraulic fracturing treatment can be traced back to the development of the Barnett shale by Mitchell Energy. After experimenting with multiple stimulation techniques, Mitchell Energy found success by combining vertical wells with hydraulic fracturing in the late 1990s. A few years later the development of the Barnett shale was evident and is illustrated by the natural gas production rate that went from 216 MMcf per day in 2000 to 3,045 MMcf per day in 2007 (RRC 2016a). The lessons and techniques learned in the Barnett shale were used in other unconventional plays like the Bakken, Eagle Ford, Haynesville, Marcellus, Niobrara, Permian and Utica. These plays accounted for more than ninety percent of the domestic oil and gas production growth between 2011 and 2014 and in 2015 produce 4 MMbbl of oil per day and more than 44,127 MMcf of natural gas per day (EIA 2016).
3 0 200 400 600 800 1,000 1,200 1,400 1,600 0 1,000 2,000 3,000 4,000 5,000 6,000 2008 2009 2010 2011 2012 2013 2014 2015 O il Pr od u ct ion , M b b l/D A Y Dr ill in g P e rm its a n d Na tu ra l G as Pr od u ct ion , M M cc f/d ay
Driling Permits and Daily Production in the Eagle
Ford Shale
Oil Natural Gas Drilling Permits
The first well to obtain commercial production from the Eagle Ford shale was drilled in 2008 by Petrohawk, now a part of BHP Billiton. The well was drilled in La Salle County and flowed at 7.6 MMcf of natural gas per day from a 3,299 ft lateral well with 10 hydraulic fracturing stages. After the discovery of the play the interest and activity increased. In 2008 there were 26 drilling permits for the Eagle Ford shale and in 2014 there were 5,613 drilling permits. Amid the recent decrease in hydrocarbon prices in 2015 there were 2,315 new drilling permits. Production went from less than 500 bbl of oil in 2008 to more than a million barrels of oil in 2015 (January-November data for 2015). In the same manner gas production went from 2 MMcf in 2008 to 5,366 MMcf in 2015 (January-November data for 2015) (Fig. 1.1).
Fig. 1.1 ‒ Drilling Permits and Average Daily Oil and Gas Production in the Eagle Ford Shale (RRC 2016b).
4 1.2.2 Fracture Conductivity
Fracture conductivity (𝐶𝑓) is defined as the product of fracture permeability (𝑘𝑓) and fracture width (𝑤𝑓). Fracture conductivity, Eq. 1-1, can be calculated by combining well testing analysis with production data and also by laboratory experiments,
𝐶𝑓 = 𝑘𝑓𝑤𝑓……….…………....…..…..………...… (1-1)
Cinco-Ley et al (1978) developed a mathematical model to correlate well production via an equivalent skin factor (𝑠𝑓) created by the hydraulic fracturing to the dimensionless fracture conductivity (𝐶𝑓𝑑) which is defined by Eq. 1-2. A graphical interpretation of this relationship is shown in Fig. 1.2.
𝐶𝑓𝑑 = 𝑘𝑓𝑤𝑓
𝑘𝑚𝑥𝑓……….………...…….….. (1-2)
Fig. 1.2 ‒Equivalent Skin Factor, Fracture Half Length and Dimensionless Fracture Conductivity (Cinco-Ley et al. 1978).
5
Where 𝑥𝑓 is the fracture half length; 𝑘𝑓 is the fracture permeability; 𝑤𝑓 is the fracture width and 𝑘𝑚 is the reservoir permeability. If fracture half-length and production data are known the fracture conductivity can be calculated from this model. The model was developed for an isotropic homogenous reservoir bounded by two impermeable zones. The model assumes the reservoir fluids are slightly compressible and with constant properties. Fluids are produced through a vertically fractured well.
Determination of fracture conductivity via laboratory experiments has the advantage of controlling and measuring the testing conditions which in turn reduces the unknown variables. The conductivity of a proppant pack in the laboratory is measured following the procedure outlined by ISO 13503-5:2006. This procedure consists of flowing 2% KCL solution thorough a proppant pack loaded between two smooth surfaces, commonly sandstone, and the pressure drop along the flow path is recorded. The process is repeated at multiple closure stresses, each closure stress is maintained for fifty hours. By following this procedures multiple labs can obtain comparable results regarding the quality of the proppants.
There are many variations of modified ISO tests used by the different laboratories. The most common variations are to increase the sample’s thickness in order to measure fluid leak off, maintaining the closure stress for relatively short periods and using rough fracture faces (Zhang 2014; Aweloke 2013). Another variation is flowing dry gas, wet gas, fresh water, brines individually or in multiphase flow instead of the 2% KCL brine (Barree et al. 2009; Ramurthy et al. 2011).
6
1.2.3 Fracture Conductivity Controlling Parameters
Due to complex fracture geometries and proppant settling effects it is likely that the proppants are not distributed uniformly and that the fractures contained in the stimulated area may have no proppant, low concentration of proppants and/or high proppant concentration (Warpinski 2009). It has been reported by multiple authors that fractures with a partial monolayer can have a higher conductivity than fully packed fractures at low closure stress (Brannon et al. 2004; Darin and Huitt 1959; Parker et al. 2005; Zhang. 2014). The explanation for this phenomena is that partial monolayer will have a higher porosity than the packed proppants which will yield a higher flow area. However, the small fracture width will create a high fluid velocity which will increase the non-Darcy effect. This problem will become more severe as closure stress increases and the fracture width reduces as a consequence. Furthermore, the load will concentrate into a few proppants which results on a higher stress per proppant. Brannon et al. 2004 presented Eq. 1-3 to calculate the required proppant concentration to obtain a full proppant monolayer.
𝐶𝑎 = 5.20(1 − ∅)𝛾𝑝𝑑𝑝 ………...…………...…...…... (1-3)
Where 𝐶𝑎 is the minimum proppant concentration required to obtain a full monolayer in lbm/ft2; ∅ is the minimum obtainable porosity of the proppant pack in fraction; 𝛾𝑝is the proppant specific gravity in g/cm3 and 𝑑𝑝is the average proppant
diameter in inches. Gillard et al. (2010) presented a hydraulic fracturing design that combines a special pumping schedule and fluids additives to create a proppant
7
configuration composed of proppant pillars and channels (Fig. 1.3). Fracture conductivity was measured with a short term conductivity experiment following the API RP 61, 1989 procedure. The conductivities obtained were between 1.5 to 2 orders of magnitude higher than the theoretical maximum that can be obtained with a proppant pack. This new treatment was applied in seven wells in Argentina’s La Loma La Lata field, low permeability sandstone field, the results were compared to eight offset wells located in the field that were fractured with a conventional method. The new method produced 53% higher initial production and a 15% increase in projected hydrocarbon recovery.
Fig. 1.3 ‒Proppant Arrangement with Proppant Pillars and Channels on a Fracture Conductivity Sample (Gillard et al. 2010).
Huitt and McGlothlin (1958) performed experiments to determine the embedment of Ottawa sand, 10-12 mesh size, in core samples where embedment is possible. They concluded that embedment of proppant is possible and as a result this diminishes the ability of the formation to maintain a propped fracture. Their equation is valid for
8
formations where the sands embed completely and no proppant crushing occurs. Kamali and Pournik (2015) developed a mathematical model for rough fractures contact, and then they combined this model with a fluid flow simulator to predict fracture conductivity under different closure stress. They concluded that the conductivity decline depends on the surface profile which dictates the mechanical interaction and deformation of the asperities.
The following studies used nitrogen to calculate fracture conductivity and the proppants were place manually in the fracture surface. Kamenov (2013) executed short term fracture conductivity (propped and unpropped) experiments in Barnett shale outcrop samples. He observed that displaced natural fractures have an order of magnitude higher conductivity than aligned natural fractures due to the contact between two non-matching rough surfaces. He determined that whenever proppant was used, conductivity was proppant dominated and the surface roughness played a smaller role. Zhang (2014) performed fracture conductivity experiments on Fayetteville, Eagle Ford and Barnett outcrops samples. He concluded that fracture conductivity for unpropped samples is well correlated to rock brittleness. By performing fracture conductivity with proppant concentrations lowerthan 0.15 lbm/ft2 he demonstrated that low proppant concentrations,
including unpropped, have and can maintain fracture conductivities in the 3-30 md-ft range at 4,000 psi closure stress for the Barnett shale outcrop samples. He also observed that when multiple layers of sands were placed in the fracture surface larger sands will have a higher conductivity than smaller sands. When comparing short term (0.5 hours) versus long term conductivity measurements (fifty hours) the short term conductivities
9
were 20% higher. Briggs (2014) performed short term fracture conductivity in two different zones of the Fayetteville shale. These zones are FL 2 and FL 3. The brittleness and conductivity of zone FL 2 were found to be greater than the ones in FL 3. The results were supported with production data that shows a greater production in FL 2 than in FL 3. Jansen (2014) used the fracture conductivities obtained by Kamenov (2013) and Zhang (2014) and compared these values with the Young’s modulus and Poisson’s ratio obtained from core plugs. These core plugs were obtained from the same rocks where the conductivities samples were acquired. He concluded that surface roughness dominates the initial conductivity and that the mechanical properties dominate the conductivity decline rate. A higher Young’s modulus corresponds to a slower decline rate in conductivity and a higher surface roughness generates a higher initial conductivity. McGinley (2015) tested the propped fracture conductivity of the Marcellus shale (outcrops). He performed his experiments with samples cut in two different directions and concluded that the anisotropy of the mechanical properties decreases the fracture conductivity.
The fracture conductivity calculated at room temperature with nitrogen and without the damage caused by fracturing fluids can be used as a baseline to examine water damage to fracture conductivity. This value has to be decreased in order to take into account all the damage mechanisms. Cooke (1973) concluded that the presence of brine and/or high temperature decreases the fracture conductivity. Cooke (1975) continued his studies of fracture conductivity and concluded that residue from guar polymer, a fracture gelling agent, greatly decreased the fracture conductivity. Davies and Kuiper (1988) demonstrated the importance of multiphase flow and how this condition decreases the
10
fracture conductivity of each phase. Fine migration, proppant embedment and non-Darcy flow can reduced the fracture conductivity by more than fifty percent (Brannon et al. 2008; Palisch et al. 2007). Awoleke (2013) concluded that an aggressive flow back rate during the cleanup process is a displacement and evaporative process that significantly reduces fracture conductivity. Proppant degradation, which leads to proppant pack porosity reduction as a consequence of chemical reactions, can occur at reservoir conditions over a relatively short period of time (less than a year). This porosity reduction leads to a decrease in fracture permeability and hence a decrease in fracture conductivity (Zheng and Ellsworth 2010; Raysoni and Weaver 2012).
The results from all the authors presented in this section agree that fracture conductivity declines with closure stress. The following list presents a summary of all the factors that affect conductivity and were discussed in this section. This study analyzes the factors presented in numbers 1-6.
1. Mechanical properties: Young’s modulus, Poisson’s ratio and Brinell hardness 2. Sample orientation
3. Brittleness calculated from the mineralogy 4. Proppant concentration and type
5. Surface attributes: roughness and fracture area 6. Closure stress
7. Embedment
8. Proppant distribution 9. Flow back rate
11 10.Temperature
11.Long term exposure to stress 12.Multiphase flow and water damage 13.Gel damage
1.2.4 Eagle Ford Depositional Environment and Lithofacies
The Eagle Ford shale play extends from the San Marcos Arch to South Texas into Mexico. It was deposited during the upper cretaceous and overlies the Buda limestone and underlies the Austin Chalk. The Eagle Ford shale is limited on its west flank by uplifting that resulted on an outcrop belt. The shelf margin, formed by the Edwards reef and the Siglo reef trends, marks the eastern boundary of the Eagle Ford. To the north the San Marcos arch marks the division between the Eagle Ford shale and the Eagle Bine. On the south side the Eagle Ford shale extends into Mexico (Fig. 1.4). The thickness of this formation decreases in a south to north trend. Maximum thickness of over 600 feet is encountered in the Maverick basin and the minimum thickness of around 50 feet is present near the San Marcos arch.
The formation is commonly divided into the lower Eagle Ford and the Upper Eagle Ford. The lower member contains the highest organic content of the play and is present across the basin. On the other hand, the upper member contains a higher carbonate content and has its maximum thickness in the Maverick basin, southwest region, and thins out towards the San Marcos arch (Martin et al. 2011; Tian et al. 2012). The mineralogy and total organic content show great variation across the play. The dominant mineral is calcite
12
followed by quartz and clays (Martin et al. 2011; Tian et al. 2012). The total organic content depends on the depositional environment as well as on the maturation. There is not a well-defined trend in this parameter across the play and it seems to be dominated by regional geology and fluid maturation (Fig. 1.5).
13
Fig. 1.5 ‒ Total Organic Content of the Lower Eagle Ford Shale Calculated from Sonic and Resistivity Logs (Tian et al. 2013).
More than 500 well logs were analyzed by Tian et al. (2012) and Tian et al. (2013) in order to generate the previous two maps.
1.3 Problem Description, Objectives and Significance
It is accepted by the industry that well production in shale reservoirs is directly related to fracture conductivity. Most of the published data for conductivity measurements follow the API standard. This procedure uses sandstone samples with a smooth surface and high concentration of proppants. While this procedure is an effective way to examine the quality of the proppants, it is not representative of the fracture conductivity in shale formations for the common low proppant concentration job. Furthermore, the Eagle Ford shale is a heterogeneous formation as seen by the wide range of flow rates and fluids produced in wells located a few miles apart (Fig. 1.6). The heterogeneity was captured by a study performed by Donovan et al. (2012) where five different geological facies for the
14
Eagle Ford shale were defined. There is not much public literature that analyzes the impact that different lithologies within a formation have on fracture conductivity.
The work here focuses on the unpropped, fracture conductivity without proppants, and propped fracture conductivity of the Eagle Ford shale. Realistic proppant concentration and proppant types for this formation are used. The fracture orientation with respect to the bedding planes and the fracture surface attributes are analyzed and their impact on conductivity is discussed. Heterogeneity of the formation is addressed by testing fracture conductivity samples from the aforementioned five different facies present in the Eagle Ford shale. The mineralogy of the samples was determined in order to validate our samples relationship to the Eagle Ford shale and to analyze its impact on fracture conductivity.
Fig. 1.6 ‒ Gas Oil Ratio for Black Oil Wells in the Eagle Ford Shale (Tian et al. 2013).
The objective of this study is to develop a deep understanding of the Eagle Ford shale unpropped and propped fracture conductivity. This knowledge in turn could be used
15
to improve the understanding of the production of the Eagle Ford shale and to design the appropriate proppant type and concentration used during the stimulation process. It is worth pointing out that the samples used for this study were gathered at outcrops. With this in mind the deliverables of this project are the following:
(1) To adapt and improve the laboratory procedures to measure fracture conductivity (2) To measure unpropped and propped fracture conductivity of the five different
geological facies of the Eagle Ford shale-outcrop samples.
(3) To investigate the relationship(s) between fracture conductivity and surface attributes, fracture orientation and geological facies.
(4) To analyze the difference between propped and unpropped fracture conductivity in the Eagle Ford shale-outcrop samples.
1.4 Approach
The experimental procedures used in this work can be divided into three main sections. The first section consists of defining the experiments that will be performed. The second section involves performing the laboratory experiments in order to gather data and the last section is analyzing the data. Experiments were designed to measure the unpropped and propped fracture conductivity under realistic conditions for the five different geological facies of the Eagle Ford shale. In order to perform the experiments, samples were collected from the five different zones of the Eagle Ford shale. Three different fracture orientations are considered in this study. After the collection process the rocks are sent to Kocurek Industries where they are cut and fracture under mode I tension into modified API
16
conductivity dimensions. Fracture conductivity at different closure stresses is determined based on laboratory measurements following the procedure developed from previous studies (Kamenov 2013; Aweloke 2013; Briggs 2014; Guzek 2014; Zhang 2014; McGinley 2015). The fracture area and fracture roughness are calculated from data taken with a surface profile meter. The mineralogy of the samples is determined in order to validate our sample’s relationship to the Eagle Ford shale and to analyze the impact this parameter has on fracture conductivity. In addition to these experiments pictures of every sample are taken. Once the data is collected, the unpropped and propped conductivities as well as the proppant effect for each zone are analyzed and discussed. Conductivity dependence on fracture area, fracture roughness and mineralogy is explored. After analyzing each zone, comparisons across zones are made, then conclusions for each zone are presented and general conclusions are drawn by combining the analyses of the five different zones.
1.5 Structure of Thesis
The structure of this thesis is the following. Section 1 includes a literature review of the fracture conductivity, its controlling parameters and the Eagle Ford shale and its importance. In this section the link between these topics is made and importance of studying the fracture conductivity of the Eagle Ford shale is presented.
Section 2 introduces the samples and their relationship to the Eagle Ford shale. The experimental design and procedures are introduced as well as the methods used to calculate the fracture conductivity.
17
Section 3 focuses on the experimental results. The conductivities for each geological facies are analyzed and their relationship to the surface attributes is explored. Section 4 builds on the results presented in the previous section and uses them to create comparisons across the five different geological facies. The mechanical properties used in this study are introduced. Based upon this, the controlling parameters on fracture conductivity for the Eagle Ford shale are found.
Section 5 summarizes the results, presents the limitations of this work and provides recommendations for future work.
18
2. EXPERIMENTAL DESIGN AND METHODOLOGY
This section presents a description of the samples, laboratory design, equipment and laboratory procedures used to calculate fracture conductivity.
2.1 Sample Collection
Eagle Ford shale samples were collected at Antonio Creek and Lozier Canyon. Both outcrops are several hundred feet thick and extend for a few thousand feet laterally. The outcrops are separated by a few miles and as you move from Lozier canyon towards Antonio creek, different facies of the Eagle Ford shale are accessible for collection. These outcrops are located in West Texas (Fig. 2.1). At these locations Donovan et al. (2012) defined five different geological facies (A, B, C, D and E) each with distinct lithological characteristics and geochemical response that can be correlated to the subsurface of South Texas. In this classification, facies A lies right above the Buda limestone, followed by facies B which has the highest total organic content and production. Facies A and B are known as the lower Eagle Ford shale, whereas Facies C, D and E are also known as the upper Eagle Ford shale. Facies E underlies the Austin chalk. (Fig. 2.2).
19
Fig. 2.1 ‒ Map of the Geographic Location of Antonio Creek and Lozier Canyon (Gardner et al. 2013).
Fig. 2.2 ‒ Five Different Geological Facies of the Eagle Ford Shale with Their Corresponding Gamma Ray Response and the Industry Nomenclature for the Formation-
20
Rock samples were collected from the five different geological facies. Mineralogy of the samples via x-ray diffraction was determined by Knorr (2016). A summary of each facies as described by Donovan et al. (2012) is presented as well as the determined mineralogy in order to link the samples to the outcrops.
Facies A which lies at the bottom of the formation is formed by light gray limestone (grainstones/packstones) separated by thin calcareous mudstone beds. The mineralogy of the collected samples was found to be 84% Calcite, 7% Quartz, 3% Dolomite, 1% Chlorite, 2% Albite and 3% Kaolinite (Fig. 2.3).
Fig. 2.3 ‒ Mineralogy (Weight Percent) and Collected Rock Sample (Sharpie for Scale) – Facies A.
Facies B is a black organic rich calcareous mudstone with some interbedded limestone. The mineralogy of the collected samples was found to be 70% Calcite, 18% Quartz, 2% Dolomite, 1% Pyrite and 9% Kaolinite (Fig. 2.4).
21
Fig. 2.4 ‒ Mineralogy (Weight Percent) and Collected Rock Sample (Pen for Scale) – Facies B.
Facies C is formed by grainstone prone limestone with mudstone intervals. The mineralogy of the collected samples was found to be 77% Calcite, 11% Quartz, 1% Pyrite, 3% Chlorite, 2% Kaolinite and 5% Illite (Fig. 2.5). Facies D is a marl (calcium carbonated with some sand, silt and/or clays) with some limestone. The color of this facies is pale yellow ochre. The mineralogy of the collected samples was found to be 93% Calcite, 2% Quartz, 1% Pyrite, 1% Dolomite and 3% Kaolinite (Fig. 2.6).
22
Fig. 2.5 ‒ Mineralogy (Weight Percent) and Collected Rock Sample (Credit Card for Scale) – Facies C.
Fig. 2.6 ‒ Mineralogy (Weight Percent) and Collected Rock Sample (Pen for Scale) – Facies D.
Facies E is formed by thin interbedded limestones (grainstone) and calcareous mudstone. The color of this facies is yellow ochre. The mineralogy of the collected samples was found to be 93% Calcite, 2% Quartz, 1% Pyrite, 1% Dolomite and 3% Kaolinite (Fig. 2.7).
23
Fig. 2.7 ‒ Mineralogy (Weight Percent) and Collected Rock Sample – Facies E. The samples collected for this study match the expected mineralogy of the outcrops. The high quartz content in zone B can be explained by the positive relationship between quartz and total organic content. Miceli-Romero (2014) determined the total organic content of samples from Lozier canyon via Rock Eval Pyrolysis (Table. 2.1). Table. 2.1 ‒ Total Organic Content (Weight Percent) Samples from Lozier Canyon.
Sub-Facies B1 and B2 Combined into Sub-Facies B.
In order to link the outcrops to the subsurface, four major sequences divided by beds (K63sb, K64sb, K65sb, K70sb and K72sb) with specific well log responses were identified. (Fig. 2.8). The sequences were first identified at the outcrops and then correlated to wells in the Eagle Ford shale (Fig. 2.9).
Facies TOC, % A 2.8 B 5.5 C 1.1 D 0.6 E 0.1
24
Fig. 2.8 ‒ Biostratigraphic, Geochemical, Petrophysical and Lithofacies Interpretation of Lozier Canyon and Sequence Interpretation (Donovan et al. 2012).
25
Fig. 2.9 ‒ Well Log Correlation from the Maverick Basin to the San Marcos Arch Modified from Donovan et al. (2012).
The samples collected and used in this study are representative of the outcrops where they were collected. In turn these outcrops are representative and can be traced to the Eagle Ford shale in the subsurface. Therefore, the samples collected and used in this study are a good representation of Eagle Ford shale.
26
2.2 Sample Preparation for Fracture Conductivity Test
Fractures propagate perpendicular to the minimum principal stress. On a typical sedimentary basin the overburden is the maximum principal stress and the other two principal stresses are located on the horizontal plane. Therefore a fracture will propagate on a plane perpendicular (fracture plane) to the minimum principal horizontal stress. The fracture plane will cut across bedding planes and when a complex fracture network is created, fractures along the bedding are needed to connect the different fracture planes. In order to represent these orientations, samples in three directions with respect to the bedding plane were obtained (Fig. 2.10).
27
The rock samples were cut and fractured into modified API conductivity cell dimensions (7 inches long, 1.65 inches wide and at least 3 inches thick) by Kocurek industries. Sandstone inserts were used to reach the 6 inches in thickness required for the modified API conductivity cell (Fig. 2.11).
Fig. 2.11 ‒Modified API Conductivity Sample Dimensions - (Kamenov 2013).
The sample orientations represent three principal flow directions; two in the vertical direction with one parallel to the bedding plane (X0) and another one perpendicular to the bedding plane (X90) as well as one in the horizontal direction (Z). (Fig. 2.12). X0 and X90 represent flow in the fracture plane and therefore are more relevant on the Eagle Ford shale. For this reason samples in the X0 and X90 direction were obtained for the five different geological facies. Samples in the Z directions were obtained only for zone B since it is the facies with the highest total organic content and production.
28
Fig. 2.12 ‒ Sample Orientation: Parallel (X0) and Perpendicular (X90) to the Bedding Plane and on the Bedding plane (Z) Modified from McGinley 2015. 2.3 Proppant Concentration
For this study proppants were placed manually in the fracture face. The proppants used were manufactured in compliance with ISO 13503-2. Proppant concentration is expressed in lbm/gal (Ppg) and also in lbm/ft2. In this study the conversion of the two units, lbm/gal and lbm/ft2 was performed by first calculating the fracture area with Eq. 2-1. Sample calculation is shown in Eq. 2-2.
29 𝐴𝑓𝑟𝑎𝑐𝑡𝑢𝑟𝑒= ((7 𝑖𝑛 − 1.65 𝑖𝑛) × 1.65 𝑖𝑛) + 𝜋 × 1.65 𝑖𝑛2 4 = 10.97 𝑖𝑛2 𝑓𝑡2 144 𝑖𝑛2= 0.076 𝑓𝑡2 ... (2-2)
Then we calculate the weight of the proppants that will be used for a given surface concentration (in this case 0.1 lbm/ft2) with Eq. 2-3. Sample calculation is shown in Eq. 2-4. 𝑃𝑟𝑜𝑝𝑝𝑎𝑛𝑡 𝑚𝑎𝑠𝑠 = 𝐷𝑒𝑠𝑖𝑟𝑒𝑑 𝑐𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 𝑙𝑏𝑚 𝑓𝑡2 × 𝐴𝑓𝑟𝑎𝑐𝑡𝑢𝑟𝑒... (2-3) 𝑃𝑟𝑜𝑝𝑝𝑎𝑛𝑡 𝑚𝑎𝑠𝑠 =0.1 𝑙𝑏𝑚 𝑓𝑡2 × 0.076 𝑓𝑡 2 = 0.0076 𝑙𝑏𝑚... (2-4)
To convert a proppant mass per unit area concentration to a proppant mass per unit volume, Ppg, a fracture width needs to be assumed. With this assumption and Eq. 2-5 the proppant concentration per unit volume is converted from the units of lbm/ft2. For this sample calculation we assume a fracture width of 0.15 in. Sample calculation is shown in Eq. 2-6.
𝐶𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛/𝑣𝑜𝑙𝑢𝑚𝑒 = 𝑃𝑟𝑜𝑝𝑝𝑎𝑛𝑡 𝑚𝑎𝑠𝑠
30 𝐶𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 𝑣𝑜𝑙𝑢𝑚𝑒 = 0.1 𝑙𝑏𝑚 𝑓𝑡2 ×0.15 𝑖𝑛× 𝑓𝑡 12 𝑖𝑛 = 8𝑙𝑏𝑚 𝑓𝑡3 𝑓𝑡3 7.48052 𝑔𝑎𝑙 = 1.069 𝑙𝑏𝑚 𝑔𝑎𝑙... (2-6) As discussed the equivalent Ppg concentration depends on the fracture width assumed. Table. 2.2 presents the equivalent proppant concentration per unit volume (Ppg) for multiple combinations of assumed fracture widths (in) and proppant concentrations per unit area (lbm/ft2). As discussed by Zhang (2014) the assumed fracture widths are at the end of the pumping stage and ignores proppant settling before the fractures closed on the proppants.
Table. 2.2 ‒ Equivalent Proppant Concentrations (Ppg) for Multiple Combinations of Fracture Widths (in) and Proppant Concentrations per Unit Area lbm/ft2.
As a reference a 100 mesh sand grain in average has a diameter of 0.0059 in. In order to determine the proppant concentration that would be used in this study information from more than three hundred wells from fifty six different companies was analyzed. The data was obtained from drillinginfo (www.drillinginfo.com) and contains wells targeting the Eagle Ford shale in sixteen different counties in Texas (Fig. 2.13).
Width 0.10 in 0.15 in 0.20 in 0.25 in 0.4 0.02 0.04 0.05 0.06 ppg 1.2 0.08 0.11 0.15 0.19 𝑙𝑏𝑚/ 𝑓𝑡2 1.6 0.1 0.15 0.2 0.25
31 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0 5 10 15 20 25 30 35 40 45 50
C
um
ul
a
ti
v
e,
%
F
requency
Concentration, ppg
Fig. 2.13‒ Proppant Concentration (Ppg) Histogram for 353 Wells Located in the Eagle Ford Shale in 16 Different Counties in Texas- Data Obtained from
www. Drillinginfo.com.
More than 80% of the analyzed wells contain concentrations between 0.4 Ppg and 1.6 Ppg. The mean, median and mode are also contained in this range. We decided to use a 0.1 (lbm/ft2) because this concentrations will put the concentration in the 0.4-1.6 Ppg range for multiple fracture width assumptions (Table. 2.2). 100 mesh and 30/50 mesh proppants were chosen as the proppants for the experiments in this study. Both proppants are among the most used in the Eagle Ford shale according to the data analyzed (Fig. 2.14). Furthermore, 100 mesh is commonly used during slick-water treatment which are common in natural gas wells. In addition, previous students from our research group
32 0 10 20 30 40 50 60 70 100 40/70 Other 20/40 30/50
F
requency
Proppant Tyoe
(Zhang 2014; Guzek 2014) used these proppants and concentrations which allows result comparisons.
Fig. 2.14‒ Proppant Type Histogram for 284 Wells Located in the Eagle Ford Shale in 16 Different Counties in Texas- Data Obtained from www. Drillinginfo.com.
The data (www.drillinginfo.com) used to select the proppant type and concentration was not reported in a consistent manner, in order to use this data the following process was used. All the proppant types used in each well were recorded. For the proppant concentration: If the data was reported as an average Ppg using a single proppant for the job it was used as reported. If multiple proppant concentrations and proppants were used and reported a weighted average was calculated and the result was used. If the total proppant mass and total fluid volume were reported, the concentration was calculated by dividing the total mass over the total volume of fluids used. If the data
33
was reported in a different manner than the ones previously discussed it was not used. We recognize that when the proppants are added into the slurry mix they increase the volume of the mixtures and therefore they change the proppant per unit volume concentrations. However, it was found that this change is insignificant for the concentrations used in this study. This point is illustrated with sample calculations for a 1 Ppg concentration using 2.65g/cm3 = 165.43 lbm/ft3(density of quartz) for proppant grain density. If the proppant volume is ignored Eq. 7 has to be used. Sample calculation is shown in Eq. 2-8. Otherwise Eq. 2-9 is used. Sample calculation is shown in Eq. 2-10.
𝐶𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑖𝑜𝑛/𝑣𝑜𝑙𝑢𝑚𝑒 = 𝑃𝑟𝑜𝑝𝑝𝑎𝑛𝑡 𝑀𝑎𝑠𝑠 𝑉𝑓𝑙𝑢𝑖𝑑 ... (2-7) 𝐶𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑖𝑜𝑛 𝑝𝑒𝑟 𝑢𝑛𝑖𝑡 𝑣𝑜𝑙𝑢𝑚𝑒 = 1 𝑙𝑏𝑚 1 𝑔𝑎𝑙... (2-8) 𝐶𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 𝑝𝑒𝑟 𝑢𝑛𝑖𝑡 𝑣𝑜𝑙𝑢𝑚𝑒 = 𝑃𝑟𝑜𝑝𝑝𝑎𝑛𝑡 𝑀𝑎𝑠𝑠 𝑉𝐹𝑙𝑢𝑖𝑑+𝑉𝑝𝑟𝑜𝑝𝑝𝑎𝑛𝑡……..…...… (2-9) 𝐶𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 𝑝𝑒𝑟 𝑢𝑛𝑖𝑡 𝑣𝑜𝑙𝑢𝑚𝑒 = 1 𝑙𝑏𝑚 1 𝑔𝑎𝑙+ 1 𝑙𝑏𝑚 165.43 𝑙𝑏𝑚 𝑓𝑡3 𝑓𝑡3 7.48 𝑔𝑎𝑙 = 0.96𝑙𝑏𝑚 𝑔𝑎𝑙... (2-10)
The previous calculations were repeated for different proppant concentrations (ppg) with different assumptions for the proppant grain densities (Fig. 2.15).
34
Fig. 2.15 ‒ Proppant Concentration (Ppg) Ignoring and Taking into Account the Proppant Volume for Multiple Proppant (Grain) Densities.
The proppant volume effect on the ppg concentration increases as the concentration per volume increases and as the proppant grain density decreases. From the graph we can see that for normal sands used during slick-water treatments the proppant volume effect on the concentrations is insignificant for the concentrations represented in this study. Therefore, dividing the total mass over the total fluids used in a treatment to calculate the concentration does not introduce a significant error.
2.4 Laboratory Procedures
This section describes the apparatus, experimental design and procedures used for fracture conductivity measurements. The steps followed in this work can be seen in Fig.2.16.
0.00 0.50 1.00 1.50 2.00 2.50 0.00 0.50 1.00 1.50 2.00
Ig
no
ri
ng
P
ro
ppa
nt
V
o
lum
e,
ppg
Considering Proppants Volume, ppg
2 g/cm^3 1 g/cm^3 4 g/cm^3 Concentrations Represented in
35
Fig. 2.16 ‒ Process Flow Chart to Calculate and Analyze the Eagle Ford Shale Fracture Conductivity.
2.4.1 Apparatus
Fracture conductivity is determined via laboratory experiments. The main laboratory equipment components used in this study to calculate fracture conductivity and to characterize the surface attributes are the following.
Hydraulic Load Frame
Modified API Conductivity Cell
Sample Gathering
Five Different Geological Facies
Experimental Design
Fluids, Proppant and Conductivity Direction
Sample Cutting
Modified API Dimensions 3 Directions
Sample Preparation
Profile-meter, Epoxy and Teflon Tape
Fracture Conductivity Experiments
Unpropped, 100 Mesh and 30/50 Mesh-At Different Closure Stresses Data Processing and Analyzing Fracture Conductivity, Surface Attributes & Geological characteristics
36 Mass Flow Meters
Pressure Transducers Surface Profile-Meter
The load frame used is a GCTS FRM4-1000-50S four column vertical standing frame. It has a 1,000 kN static compression load capacity and a 800 kN dynamic load capacity. The stroke length is 5 mm and the maximum velocity is 80 mm per minute. As reported by Kamenov (2013) the modified API cell is made of stainless steel and consists of a cell body, two side pistons and two flow inserts (Fig. 2.17). The cell body dimensions are 10 inches long, 3.25 inches wide and 8 inches in height. It can accommodate a core sample 7 inches long, 1.65 inches wide and up to 7 inches thick. Three pressure ports are located in the middle of the cell body. The pistons are 7 inches long, 1.65 inches wide and 3 inches in height and both have a viton polypak seal (O-ring) to prevent leakage. The pistons also have two drilled holes (can be covered) that can be used to simulate fluid leakage if desired. The flow inserts also have viton polypak seal (O-ring) and connect the upstream and downstream flow lines to the cell body.
37
Fig. 2.17 ‒ Modified API Conductivity Cell Modified from Kamenov 2013.
The mass flow meters used were thermal mass flow controllers manufactured by AALBORG. Two flow meters were used. For high flow rates a mass flow meter with range from 0 to 10 standard liters per minute was used. For low flow rates a mass flow meter with range from 0 to 1 standard liters per minute was used. Both flow meters have an accuracy of ± 1% of their maximum flow rate. The maximum working gas pressure is 1,000 psi and the optimum working gas pressure is 20 psi (Fig. 2.18). Two pressure transducers were used during the experiments, both were Validyne DP 15 (Fig. 2.19). One measures the pressure inside the conductivity cell and the other measures differential pressure along the flow path (fracture). The diaphragms can be easily changed to account for different pressure drops expected during the experiment. The transducers have a ± 2.5% accuracy of their full pressure range. A detailed procedure for the calibration process
38
can be found in Awoleke’s work (Aweloke 2013). Further details on troubleshooting the transducers have been presented by Zhang (2014) and McGinley (2015).
Fig. 2.18 ‒ Low and High Flow Rates Mass Flow Meters.
39
A schematic of fracture conductivity experimental setup can be seen in Fig. 2.20. The fracture faces were scanned with a surface profilometer and the data obtained was used to calculate the fracture area and roughness (Fig. 2.21).
Fig. 2.20 ‒ Schematic of Fracture Conductivity Experimental Setup.
40 2.4.2 Experimental Design
The number of experiments performed was limited by the amount of rock samples collected and/or their dimensions. We obtained as many conductivity samples as possible from the gathered rock samples. The plan was to first measure the unpropped fracture conductivity follow by the conductivity for a 100 mesh proppant concentration of 0.1 lbm/ft2, and to finish by measuring the conductivity using 30/50 proppant with a concentration of 0.1 lbm/ft2. This way three tests can be conducted on one sample. Table.
2.3 list the original designed experimental conditions for this study.
Table. 2.3 ‒ Conductivity Samples Obtained from the Collected Eagle Ford Shale Outcrop Rocks. Geological Facies Orientation Number of Samples Unpropped 0.1 lbm/ft2 -100 Mesh 0.1 lbm/ft2 -30/50 Mesh Number of Experiments E X0 2 2 2 2 6 X90 2 2 2 2 6 D X0 2 2 2 2 6 X90 2 2 2 2 6 C X0 2 2 2 2 6 X90 3 3 3 3 9 B X0 3 3 3 3 9 X90 4 4 4 4 12 Z 3 3 3 3 9 A X0 2 2 2 2 6 X90 2 2 2 2 6 Total --> 27 26 26 26 81
41
Some samples were damaged and could not be reuse after the first conductivity experiment. There are also samples that supported two conductivity experiments but could not be used for a third one (Fig. 2.22). A small number of samples were able to sustain the three conductivity experiments.
Fig. 2.22 ‒ Damaged Sample from Facies B After 100 Mesh Experiment.
In addition to sample destruction, another limitation was that some unpropped test required a high differential pressure between the cell and the ambient in order to obtain flows within the range of the used mass flow meters. In some cases the differential pressure required was greater than the range of the pressure transducer calibration tool used (30 psi). In these cases unpropped conductivities could not be measured. Taking into consideration these limitations, a summary of the actually performed experiments is presented in Table. 2.4.
42
Table. 2.4 ‒ Performed Fracture Conductivity Experiments for the Eagle Ford Shale Samples Geological Facies Orientation Number of Samples Unpropped 0.1 lbm/ft2 -100 Mesh 0.1 lbm/ft2 -30/50 Mesh Number of Experiments E X0 2 2 2 1 5 X90 2 1 2 2 5 D X0 2 2 2 1 5 X90 2 1 2 1 4 C X0 2 1 2 0 3 X90 3 2 3 2 7 B X0 3 2 3 1 6 X90 4 2 4 3 9 Z 3 3 3 1 7 A X0 2 2 1 0 3 X90 2 2 2 1 5 Total --> 27 20 26 13 59
2.4.3 Methodology of Surface Roughness and Surface Area Measurements
The first step taken after receiving the conductivity samples from Kocurek industries was to label them for future references. Samples names have the following format: first letter of formation name for example E for Eagle Ford followed by the geological facies (A-E), followed by an underscore and the sample orientation (X0, X90 and Z), followed by another underscore and the sample number. The experimental files have the experiment type added to the sample name at the end after an underscore (U for unpropped, 100 for 0.1 lbm/ft2 100 mesh and 30/50 for 0.1 lbm/ft2 30/50 mesh). For example sample
43
EB_X0_1_U is a sample from the Eagle Ford shale that represents flow in the X0 direction, parallel to the bedding plane, which is the first sample with these conditions and the experiment was run with no proppants. Next, pictures of the samples were taken. After these steps are completed the samples are ready to be scan.
The samples are placed on the profilometer (Fig. 2.21) where both fracture surfaces are scanned. The apparatus has a full scale resolution of 1 inch and an accuracy of ±0.000001 inches. The scanned dimensions are 7 inches long, 1.7 inches wide and the data is collected every 0.05 inches. The profilometer generates a text file with three columns: x position, y position and z reading. This file is the input to a MATLAB code where the fracture roughness and area are calculated. The code only uses data points collected in a rectangular area in the middle of the sample. This is to remove the edges of the sample where the profilometer readings are not reliable. For comparison the area of the rectangle where data is used is 8.81 𝑖𝑛2 and the area of the conductivity cell is 10.97 in2. Surface roughness was calculated with Eq. 2-11 and Eq. 2-12 via the root mean square method (RRMS). 𝑅𝑅𝑀𝑆 = √1 𝑛∑ 𝑧𝑖 2 𝑛 𝑖=1 ……….………... (2-11) 𝑧𝑖 = |𝑧 − 𝑧̅|………..………... (2-12)
44
Where 𝑧̅ is the average reading of all the points measured by the profilometer (inside the rectangle) and 𝑧 is the profile-meter reading at a specific point. The fracture area was calculated by dividing the rectangular area into parallelograms and using the cross product to calculate the area of each parallelogram with Eq. 2-13.
𝐴𝑟𝑒𝑎 𝑜𝑓 𝑃𝑎𝑟𝑎𝑙𝑙𝑒𝑙𝑜𝑔𝑟𝑎𝑚 𝐹𝑜𝑟𝑚𝑒𝑑 𝑏𝑦 𝑥⃗ 𝑎𝑛𝑑 𝑦⃗ = |𝑥⃗ × 𝑦⃗|…... (2-13)
The area of all the parallelograms were added in order to calculate the rectangular area (fracture area). The process to generate the text file from the profilometer file was presented by (McGinley 2015). The output of the MATLAB code is shown in Fig. 2.23. Fracture area and roughness were calculated for all samples.
45 2.4.4 Methodology of Sample Preparation
In order to create a good seal between the conductivity sample and the modified API conductivity cell an epoxy (Momentive RTV 627) was applied to the conductivity samples. The epoxy was applied in a mold 0.1 inches wider than the conductivity sample and 0.003 inches wider than the modified API conductivity sample (Fig. 2.24). This creates a tight fit between the conductivity sample and the modified API conductivity which makes significant leakages unlikely even at high pressures. The procedure to apply the epoxy and the common mistakes during the process have been covered by previous studies. (Kamenov 2013; Briggs 2014; Guzek 2014; Zhang 2014; McGinley 2015).
Fig. 2.24 ‒ Aligned Conductivity Sample in the Sample Preparation Mold (Guzek 2014).
The only part of the sample preparation procedures that have not been discussed before is how to remove the conductivity cell from the sample preparation mold. This
46
procedure starts after the sample have been taken out of the oven for a few hours and the mold is at room temperature.
(1) Remove the mold baseplate and all the screws (4) of the sample preparation mold (2) Place the sample preparation mold under a hydraulic press
Note: there should be no obstruction at the location of the conductivity sample (3) Gently apply force with the hydraulic press to the conductivity sample, this
should open the preparation mold and release the sample
Fig. 2.25 ‒ Removing the Conductivity Sample Coated With Epoxy from the Sample Preparation Mold.
(4) Cut the epoxy to expose the flow inlet and outlet flow ports as well as the three pressure ports. A detailed procedure for this was presented by McGinley (2015)
47
(5) Use a high density gas line Teflon tape to isolate the pressure ports and minimize leakage. Apply two lines of Teflon tape below the pressure ports, two lines above the pressure ports and two lines in between the pressure ports.
Note: Each Teflon tape line is formed by three wraps of the tape around the conductivity sample
(6) Apply vacuum grease in the spaces between the Teflon tape lines Note: Do not apply vacuum grease near the pressure ports
(7) The Sample is ready to be placed in the Modified API conductivity sample for unpropped experiments. Briggs (2014) described this process.
Fig. 2.26 ‒ Conductivity Sample Ready to be Inserted into the Modify API Conductivity Cell for Measurements.
48
For propped conductivity experiments an extra step involving the proppant placement has to be taken.
(1) Using the digital scale, measure the desired proppant mass for the experiment (2) Cut a line in the epoxy connec
